为了模拟人类视觉系统成像机制, 真实再现人眼对外界场景的视觉感知, 为机器视觉提供符合生物视觉特性的理想图像,从色度学的角度研究了彩色图像的真实影像再现技术。针对色貌模型存在图像模糊、对比度低的缺点, 在iCAM色貌模型的基础上, 引入了人工神经网络的非线性Sigmoid函数, 提出了基于Sigmoid-iCAM色貌模型的真实影像再现算法, 并给出了该模型处理不同类型图像的控制参数统计规律。与基于CIECAM02、iCAM色貌模型的真实影像再现算法相比, 该算法在对色调失真图像实现颜色复原的同时, 有效地提高了图像的对比度和局部细节, 很好地实现了视觉系统的颜色恒定性。

In order to simulate the imaging mechanism of human vision system, reproduce the same scene perception of human eyes and provide machine vision the ideal images which are in accordance with biological visual characteristics, color image reproduction technology is studied from the colorimetry . Aiming at the shortcomings of image color appearance model fuzzy and low contrast,the visual local adaptation mechanism based on the Sigmoid function of artificial neural network is introduced, then, a realistic image rendition algorithm based on Sigmoid-iCAM color appearance model is proposed. Compared with the realistic image rendition algorithms based on CIECAM02 and iCAM models, this algorithm can improve the image contrast and local details effectively , and can achieve a good color restoration for color images with color distortion, and can well realize the visual system of color constancy.